lar2par#
- class diffsptk.LogAreaRatioToParcorCoefficients(par_order: int)[source]#
See this page for details.
- Parameters:
- par_orderint >= 0
The order of the PARCOR coefficients, \(M\).
- forward(g: Tensor) Tensor [source]#
Convert LAR to PARCOR.
- Parameters:
- gTensor [shape=(…, M+1)]
The log area ratio.
- Returns:
- outTensor [shape=(…, M+1)]
The PARCOR coefficients.
Examples
>>> g = diffsptk.ramp(1, 4) * 0.1 >>> lar2par = diffsptk.LogAreaRatioToParcorCoefficients(3) >>> k = lar2par(g) >>> k tensor([0.1000, 0.0997, 0.1489, 0.1974])